Reasonable stochastic model and function model are the premise of accurate velocity determination, especially in the time-differenced carrier-phase (TDCP) method. This paper presents, first, an elevation-dependent stochastic model (ESM), and then gives a simplified and unified Galileo/GPS combined TDCP function model, where the inter-system bias (ISB) variations are analysed based on correlation coefficients and the scaled sensitivity matrix. To evaluate the performance of the proposed models, datasets collected at 10 multi-GNSS experiment (MGEX) stations and a vehicle kinematic experiment are employed. The results indicate that the ESM model can improve the accuracy of the velocity solution, especially for the Galileo/GPS combined system, in comparison with the equivalent weight ratio method. In contrast to the Galileo-only velocity solution, the Galileo/GPS combined velocity solution can bring improvements of about 1–1⋅5 mm/s, 0⋅5 mm/s and 1⋅5–2⋅5 mm/s in East, North and Up components, respectively. Compared with the traditional Galileo/GPS TDCP model, the simplified and unified model shows no obvious differences in all components in the environment with more visible satellites, but it performs better in a challenging environment with few visible satellites.